Symposium Organizers
Yasuhiko Ishikawa, The University of Tokyo
Brian Corbett, Tyndall National Institute
Juejun Hu, Massachusetts Institute of Technology
Shinichi Saito, University of Southampton
Chee Hing Tan, The University of Sheffield
EP7.1: Si Photonics—State-of-the-Art
Session Chairs
Tuesday PM, March 29, 2016
PCC North, 200 Level, Room 222 A
2:30 PM - *EP7.1.01
Photonic Integration on the Silicon Platform: Past, Present and Future
Lionel Kimerling 1
1 MIT Cambridge United States,
Show AbstractThe number of photonic components deployed in systems is increasing. The contribution of photonics to system cost is becoming significant. If a common manufacturing platform is shared across the industry, one can expect that cost reduction will scale with cumulative production. Integrated Silicon Microphotonics is today the only platform capable of high volume production (>10 million units) and high levels of integration. Silicon Microphotonics was an interesting research challenge; it is today a central factor in manufacturing, performance and business; tomorrow, Silicon Microphotonics is the “future-proof” solution. As system architectures continue to migrate to distributed topologies for energy efficiency, connectivity and electronic-photonic synergy assume prime importance.
Silicon Microphotonics must provide a path for the continued scaling of component performance with a standard, modular platform. The success metrics for the silicon platform are increases in yield, reliability, density; and reductions in cost, time to market, power dissipation and functional latency. The major manufacturing roadblocks to this future are design, assembly, packaging and test. Monolithic electronic-photonic integration is a universal solution, but current incompatibilities of device size, process node, thermal robustness and reliability must be rationalized.
3:00 PM - *EP7.1.02
Microring Resonators and Silicon Photonics
Fernando Manzano 1,Lorenzo Pavesi 1
1 Nanoscience Laboratory, Department of Physics University of Trento Povo Trento Italy,
Show AbstractIn this talk I will review our present work on the applications of microring physics to silicon photonics. I will details the use of microrings to:
1. measure the optical analogue of the Lamb-shift of electronic states via a vertical coupling approach
2. spontaneously generate correlated photon pairs via parametric processes on a wide spectral range
3. detect low level of contaminants in milk in a compact micro-system
4. generate chaotic sequences of optical pulses from a CW beam
3:30 PM - EP7.1.03
Co-Integrated Ge and Si Photodiodes by Direct Wafer Bonding
Nannicha Hattasan 1,Mary White 1,Alan Blake 1,Anne-Marie Kelleher 1,Brendan Roycroft 1,Brian Corbett 1
1 Tyndall National Institute Cork Ireland,
Show AbstractNear-infrared (NIR) imaging is important for applications such as telecommunications, gas sensing and medical imaging. InGaAs is an excellent direct bandgap absorbing material but it is limited to small wafer sizes and is expensive. Ge is an appropriate alternative material because it has a bandgap at ~1.7 microns and is compatible with Si-CMOS processing. Thus, the co-integration of Ge with Si photodiodes can be envisaged for use in dual band camera systems. Ge can be integrated onto Si through chemical vapor deposition and, by employing re-crystallization, low defect densities can be obtained. These integrated Ge photodiodes however have limited responsivity for surface-normal incident light at wavelengths >1300nm due to the typically thin (~1mm) Ge layer. Here, we use wafer bonding of high quality Ge wafers to Si wafers pre-prepared with photodiodes. Arrays of Si photodiodes are fabricated in shallow (400nm deep) recesses on a 100mm diameter wafer. The p-type region is formed by ion implantation of B while the n-type region is formed by diffusion of P on the upper surface. Following removal of all oxides the Si wafer is prepared for bonding using standard RCA cleans leaving the surface hydrophilic. The p-type Ge wafer is cleaned using an ammonia solution (NH4OH:H2O2:H2O 3:1:1250) for 10 minutes followed by dilute HF (HF:H2O 1:1150) for 3 minutes. After cleaning, the Ge and Si wafers are introduced into the bonding chamber which is evacuated (1×10-5 mbar). Following a remote O2 plasma activation the wafers are brought in contact at room temperature with a force of 1000N for 5 minutes. The temperature is then ramped to 100°C. The bonded wafers are kept under a 500N force at 100°C for the next hour. The wafer pair is then annealed at 200°C for 24 hrs and 300°C for another 24 hrs under a N2 environment. Transmission electron microscopy shows a thin amorphous layer of 2-5 nm thick at the interface. The Ge wafer is ground back from a thickness of 165µm to ~10µm using lapping and polishing. The alignment marks on the Si wafer are revealed and the sample is then processed into Ge and Si photodiodes. Mesas are etched through the Ge layer using an inductively coupled plasma with a BCl3:Cl2 gas mixture. The Ge photodiodes are formed by the bonded interface between the p-Ge and the n-Si with the photocurrent been taken across this junction. Ohmic contacts are made to the p-side of the Ge mesa, the n-doped Si surface and p-doped recessed Si area. 2 dimensional arrays of Ge photodiodes are interleaved with Si photodiodes and are characterized for their spectral response. A joint spectral coverage from 400nm to 1600nm is obtained opening the way for low cost broadband imaging systems
EP7.2: Si Photonics—Advanced Devices and Materials
Session Chairs
Yasuhiko Ishikawa
Shinichi Saito
Tuesday PM, March 29, 2016
PCC North, 200 Level, Room 222 A
4:15 PM - *EP7.2.01
Applications of Magneto-Optical Materials for Si Photonics
Tetsuya Mizumoto 1,Yuya Shoji 1
1 Tokyo Inst of Technology Tokyo Japan,
Show AbstractOptical nonreciprocal devices play unique roles in photonic circuits. An optical isolator transmits light waves only in a pre-determined direction and prevents undesired waves from launching into active devices. An optical circulator is an important component for constructing highly functional photonic circuits that process counter-propagating light waves. The nonreciprocal functions of bulk devices are based on the magneto-optical Faraday effect, in which a polarization plane rotates in different directions depending on the light propagation direction. Among several magneto-optical materials, a rare earth iron garnet is the best in optical fiber communications wavelength ranges owing to its large first-order magneto-optical effect and practically low optical absorption. In waveguide devices, it is hard to employ a polarization rotation because of difficulty in taking the phase matching between two orthogonal polarizations. In order to overcome this, it is effective to employ the magneto-optical phase shift.
We developed silicon waveguide optical nonreciprocal devices based on the magneto-optical phase shift in a Mach-Zehnder interferometer (MZI) configuration. In the devices, a magneto-optical garnet CeY2Fe5O12 (Ce:YIG) is directly bonded using a surface-activated bonding technology. This enables us to use the single crystalline garnet that provides a large Faraday constant of -4500 deg/cm at a wavelength of 1550 nm. The MZI is designed so that the constructive or destructive interference occurs in the forward or backward direction, respectively, by combining a direction-dependent magneto-optical phase shift and a direction-independent phase bias. The MZI was fabricated with a 220-nm-high and 450-nm-wide silicon waveguide in an SOI wafer having a 3-μm-thick buried oxide layer. A 1.5x1.5 mm2 Ce:YIG die was bonded at 200 oC after the surface activation by a plasma generated in nitrogen. An isolation of 30 dB was obtained in a fabricated isolator with an insertion loss of 13 dB at 1550 nm. An isolation >20 dB was obtained in a wavelength range of 8 nm. Also, the temperature-independent backward transmission was demonstrated in a temperature range of 20-60 oC. This was achieved by designing an MZI so as to compensate the temperature dependence of the magneto-optical phase shift with that of the phase bias. A similar approach was applied to realize an optical circulator in a silicon waveguide. A four-port circulator operation was successfully demonstrated with an isolation of 33 dB and an insertion loss of 12 dB.
4:45 PM - *EP7.2.02
Graphene-Based Optoelectronics for On-Chip Optical Interconnects
Ren-Jye Shiue 1,Dirk Englund 1
1 Massachusetts Institute of Technology Cambridge United States,
Show AbstractGraphene has emerged as a promising material for compact, high-speed and energy-efficient optoelectronics. Here, we will present our recent implementations of graphene-based electro-optic modulators and photodetectors integrated with silicon photonic integrated circuits. By coupling graphene to an optical cavity, we have demonstrated an efficient electro-optic modulator with a response speed exceeding 1 GHz. For waveguide-integrated graphene photodetectors, we achieved a maximum responsivity of 0.36 A/W and operation speeds in excess of 40 GHz. Furthermore, due to the unique optoelectronic properties of graphene, we demonstrate a new type of on-chip broadband autocorrelator based on graphene and hexagonal boron nitride heterostructures.
5:15 PM - EP7.2.03
Er Rare-Earth Silicates: Material Synthesis and Applications to Hybrid Silicon Photonics
Hideo Isshiki 1,Fumiya Kondow 1,Takuya Sugawara 2,Yasuaki Tsuyuki 1,Ghent Nakamura 1,Kojiro Sakaguchi 1,Yousong Jiang 2
1 The Univ of Electro-Communications Tokyo Japan,2 Shincron Co. Ltd. Yokohama Japan
Show AbstractRare earth (RE) silicate containing Er, such as ErxY2-xSiO5, is a candidate for the C-band light sources and optical gain media for silicon photonics. We have investigated ErxY2-xSiO5 layered crystalline thin films as the optical gain medium. Due to the Förster energy transfer between Er ions, the excitation energy dissipation, such as the cooperative upconversion and the diffusion limited relaxation, occurs in high Er content materials. In order to obtain high optical gain, the dissipation processes caused by the energy transfer was discussed and the Er content was optimized by numerical analysis considering the energy transfer. Furthermore, enhancement of the Er 1.53µm luminescence has been observed frequently in optical materials co-doped with Er and Yb owing to the energy transfer from Yb to Er ions, because absorption cross-section of Yb3+ is about one order larger than that of Er3+ at 980nm excitation band. Further effective sensitization by Yb co-doping is expected in such high Er content crystalline films with short Er3+-Er3+ distance. From the experimental results, we will show the sensitization effect above 10 times in ErxYbyY2-x-ySiO5 crystals with the optimized ratio Yb/Er=1.
Layer-by-layer process by PLD enlarges the crystalline grain and the non-radiative transition process is reduced. Then we have attempted to utilize radical-assisted sputtering (RAS) method to form the layered complex metal oxide films for the integrated device applications. RAS technique uses a high-speed rotatable drum as a substrate holder, and a radical source and metal targets set up independently around the drum. Due to control of the drum rotation and sputtering conditions, it is possible to repeat metal-sputtering and oxidation processes with layer-by-layer accuracy. By using RAS, large and smooth ErxY2-xSiO5 layered crystalline films can be obtained. In this paper, formation of ErxY2-xSiO5 layered crystalline thin films by using RAS technique is demonstrated, and the waveguide device applications are discussed.
5:30 PM - EP7.2.04
The 980nm Quasi-Lasing Properties of Er-Y Chloride Silicate Single-Crystal Nanowires
Ye Rui 1,Xu Chao 1,Wang Xingjun 1,Zhou Zhiping 1
1 State Key Laboratory of Advanced Optical Communication Systems and Networks, School of Electronics Engineering and Computer Science, Peking University Beijing China,
Show AbstractErbium chloride silicate nanowire as a high gain amplifier material has attracted a lot attention nowadays due to its high concentration of active Er3+ ions and low propagation loss. In this study, we report on the first demonstration of 980nm quasi-lasing photoluminescence in erbium(Er)-yttrium(Y) chloride silicate nanowires pumped by 1476nm laser at a temperature of 77K (liquid nitrogen). The 980nm band emission comes from the 4I11/2 to 4I15/2 intra-4f states transition of Er3+ ions, after the cooperative upconversion of two 1480nm photons. The superb linear relationship between emission intensity and pump power presented was consistent with the characteristics of laser though the threshold point has not been captured due to the precision limit of our measurement. Due to the high crystal quality of nanowires, stark splitting in the energy state was observed. These well separated sharp emission lines within this band have a linewidth of only 0.25nm, which is the narrowest linewidth of 980nm micro-lasers to date. When the temperature was increased gradually up to 180K, the peak position and linewidth remained the same. Er/Y ratio in the nanowire was further varied to analyze the 980nm upconversion emission mechanism and a red shift of the strongest peak in the band in nanowires with higher ratio of Er/Y was observed. These novel 980nm quasi-lasing properties of Er-Y chloride silicate nanowires indicate its potential in future realization of tunable 980nm laser.
5:45 PM - EP7.2.05
Growth of Stoichiometric InP Nanowire Using Chemical Vapor Deposition
Seyed Ebrahim Hashemi Amiri 2,Sunay Turkdogan 3,Fan Fan 1,Yueyang Yu 1,Zhicheng Liu 1,Praneeth Ranga 1,Cun-Zheng Ning 1
1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States,2 Department of Chemistry Arizona State University Tempe United States,1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States,3 Electronic and Communication Engineering University of Yalova Yalova Turkey1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States
Show AbstractIII-V semiconductor nanowires have attracted considerable attention in the past few years for optoelectronic device applications such as photodetectors, solar cells, LEDs, lasers, etc. Among III-V binary compounds, InP is one of the most important semiconductors because of its direct band gap and superior properties such as high electron mobility and low surface recombination velocity. So far, growth of highly crystalline InP nanowires has only been demonstrated using MOCVD and MBE1-3. For many applications such as solar cells, the typical low pressure CVD system using simpler precursors such as InP powder have advantages of low cost. But due to the severely incongruent sublimation of InP with a much higher P-sublimation rate, such growth routes lead to significant non-stoichiometric InP nanowires. In this paper, we will present our recent approach that not only resolves this issue but also significantly extends the narrow window of growth parameters of InP to a broad range of temperature while sustaining stoichiometric InP growth. Instead of InP powder, we use elemental metallic Indium and red Phosphorous source as precursors for the growth. For the Au-catalyzed growth, our study shows that high quality InP nanowires can be grown on various substrates with different lattice mismatch (LM) such as GaAs(~3.7%), Si(~8.1%) with significantly extended range of growth temperature. The optimum growth temperature window is in the range of 500-650 C in presence of Ar+5% H2 as carrier gas and pressure of 1-10 torr using 1-1.5 nm thick Au catalyst sputtered on growth substrate. Typical growth for 30-60 min results in hexagonal shaped nanowires with diameter and length in the range of 0.5-1 um and 6-12 microns, respectively. In addition, dynamics of growth for syringe- like hexagonal NWs was statistically studied based on the interplay between the axial Vapor-Liquid-Solid (VLS) and radial Vapor-Solid (VS) growth mechanism. InP NWs grown on InP(100) showed higher vertical growth yield than those grown on Si(100) substrate, apparently due to LM difference. In all the situations, our growth has produced highly stoichiometric InP with high crystal quality. Most of our individual nanowires can sustain lasing at room temperature. We believe that our growth based on simpler, lower cost precursors and low cost CVD systems could be important for InP nanowire based solar cells, especially given the resulting high crystal quality.
References
[1] Li, Kun, et al. Nano letters (2015).
[2] Gao, Qian, et al. Nano letters 14.9 (2014): 5206-5211.
[3] Kelrich, A., et al. Nanotechnology26.8 (2015): 085303.
Symposium Organizers
Yasuhiko Ishikawa, The University of Tokyo
Brian Corbett, Tyndall National Institute
Juejun Hu, Massachusetts Institute of Technology
Shinichi Saito, University of Southampton
Chee Hing Tan, The University of Sheffield
EP7.3: Si Photonics—Emerging Devices and Sensors
Session Chairs
Wednesday AM, March 30, 2016
PCC North, 200 Level, Room 222 A
9:00 AM - *EP7.3.01
Subwavelength Engineered Optical Materials for Photonic Integration and Sensing
Pavel Cheben 1,Jens Schmid 1,Danxia Xu 1,Siegfried Janz 1,Daniel Benedikovic 1,Carlos Alonso-Ramos 1,Mohamed Rahim 1,Shurui Wang 1,Martin Vachon 1,Robert Halir 1,Gonzalo Wanguemert-Perez 1,Alejandro Ortega-Monux 1,Inigo Molina-Fernandez 1,Marie-Josee Picard 2,Michel Poulin 2,Martin Papes 1,Vladimir Vasinek 1,Milan Dado 1,Jarmila Mullerova 1
1 National Research Council Ottawa Canada,2 Teraxion Quebec Canada
Show AbstractIn this invited presentation we report our recent advances in development of subwavelength engineered materials for integrated photonics. This unique technology allows synthesis of an effective photonic medium with an unprecedented control of material properties, constituting a powerful tool for a designer of photonic integrated circuits. We have demonstrated a number of subwavelength engineered devices operating at telecom wavelengths, including fibre-chip couplers, waveguide crossings, WDM multiplexers, ultra-fast optical switches, athermal waveguides, evanescent field sensors, polarization rotators, transceiver hybrids and colorless interference couplers. The subwavelength metamaterial concept has been adopted by industry (IBM) for fibre-chip coupling and subwavelength structures are likely to become key building blocks for the next generation of integrated photonic circuits. Here we demonstrate an unprecedented control over the light coupling between optical fibers and silicon chips by constructing several types of subwavelength index engineered couplers operating at telecom (1.55 mm) and datacom (1.3 mm) wavelengths. Specifically, we report a surface grating coupler with measured fiber-chip coupling efficiency of -0.7 dB, the highest efficiency achieved to date in 220-nm silicon-on-insulator (SOI). Furthermore, we report on a broadband polarization independent fibre-chip edge coupler with a coupling efficiency exceeding 90%, on the first implementation of a subwavelength structure for laser-to-SOI chip coupling experiments with an InGaAsP/InP buried heterostructure laser operating at 1.3 mm wavelength, and on an ultra-broadband dispersion engineered multimode interference coupler designed for the wavelength range 1260 nm - 1675 nm, exceeding the O, E, S, C, L and U optical communication bands. Implementations of a subwavelength engineered optical waveguide core material in the SOI evanescent field waveguide sensor and the Fourier-transform silicon photonic spectrometer chip will also be presented.
9:30 AM - *EP7.3.02
Photonic Quantum Computing
Jeremy O'Brien 1,Damien Bonneau 1
1 University of Bristol Bristol United Kingdom,
Show AbstractOf the various approaches to quantum computing, photons are appealing for their low-noise properties and ease of manipulation at the single qubit level; while the challenge of entangling interactions between photons can be met via measurement induced non-linearities. However, the real excitement with this architecture is the promise of ultimate manufacturability: All of the components---inc. sources, detectors, filters, switches, delay lines---have been implemented on chip, and increasingly sophisticated integration of these components is being achieved. We will discuss the opportunities and challenges of a fully integrated photonic quantum computer.
10:00 AM - EP7.3.03
Scalable Integrated Optical System for Measuring Small Vibrations for Sensing Applications
Amanda Harvey 1,Raj Singh 1,Nirmal Govindaraju 1
1 Oklahoma State Univ Tulsa United States,
Show AbstractResonant frequency changes in microelectromechanical systems (MEMS) have been used to accurately measure small masses or for sensing applications. However, MEMS-based resonant sensor devices require elaborate steps for device fabrication, and packaging to drive and measure the signal response. This presentation discusses an alternative optical approach for measuring small vibrational changes in large aspect ratio wires with microscale dimensions. The system relies on a non-contact laser based sensing modality for detecting amplitude and frequency changes in vibrating microscale wires and cantilevers. The optical system is scalable making it suitable for eventual use in portable field applications such as chemical and biological sensors.
10:15 AM - EP7.3.04
Optical Coupler Devices and Materials to Implement the Hybrid Optical Interconnect Plane Architecture in Si VLSI Chip
Yoonyoung Bae 1,Donghwan Ahn 1
1 School of advanced materials engineering Kookmin University Seoul Korea (the Republic of),
Show AbstractIn spite of recent great progress in silicon CMOS-compatible photonic devices and materials, no clear solution in the electronics-photonics integration architecture platform currently exists yet. Especially, most of silicon-compatible photonic devices were designed and demonstrated on SOI wafer with thick buried oxide layer, which is not a compatible substrate with the mainstream Si CMOS products. For example, the approach to implement the photonic plane on the Si wafer surface level on a par with Si CMOS (also known as the Front-End-of-Line (FEOL) photonics) requires extensive degree of substrate modification in the photonic region and the electronic-photonics architecture interference in terms of real estate is unavoidable. On the other hand, the approach to implement the photonic plane on top of inter-layer-dielectric (ILD) layer (also known as the Back-End-of-Line (BEOL) photonics) has minimal process and architecture interference with Si VLSI chip, but the high-quality crystalline semiconductor materials such as Ge are not ready especially for high-performance active photonic devices.
In this paper, we propose the hybrid optical interconnect plane architecture where the active photonic devices are formed on the surface level of regular Si wafer and the passive photonic devices are formed on the upper BEOL plane on top of metal interconnect. In such architecture, the optical coupling scheme that will vertically transport the optical signal between the devices on the upper and lower planes with high coupling efficiency is critically important.
In this paper, we introduce the design and the experimental demonstration of the vertical ring stack coupler that has high vertical transport distance with small footprint. The stacked ring coupler shows more than 90% coupling efficiency. Also, the coupling scheme between the upper waveguide bus and the lower crystalline Ge active phonic devices formed on regular Si wafer surface is presented. For example, the optical coupling design from the upper dielectric bus waveguide to the electro-absorption type Ge modulator formed on Si wafer showed more than >80% coupling efficiency. Such designs can enable the migration of Si optical interconnect architecture from the customized-for-photonics SOI wafers to the universal regular Si wafer platform. These two optical coupling schemes between photonic devices can serve as key components that enable the hybrid optical interconnect architecture in the high-performance Si CMOS chip.
10:30 AM - EP7.3.05
Growth and Properties of Amorphous MoSi Superconductor for Single Photon Detectors
David Bosworth 1,Robert Hadfield 2,Zoe Barber 1
1 Department of Materials Science and Metallurgy University of Cambridge Cambridge United Kingdom,2 School of Engineering University of Glasgow Glasgow United Kingdom
Show AbstractAmorphous superconductors have emerged as a promising alternative material for superconducting nanowire single photon detectors (SNSPDs). Detectors based on amorphous superconductors share the higher detection efficiencies, lower jitter counts and, most importantly, faster reset times of their nitride counterparts allowing for operation at telecoms wavelengths. The reduced superconducting energy gap also results in larger hotspot formation allowing for detection of lower energy photons. The amorphous nature allows for a more homogenous film to be grown which results in a spatially uniform detection efficiency and higher internal quantum efficiency [1].
Amorphous WxSi1-x has been shown to have performance comparable or exceeding that observed in traditional NbN or NbTiN based devices. Without the use of optical cavities, detection efficiencies of up to 40% have been observed. By utilising an optical stack architecture, detection efficiencies of up to 93% have been recorded [2]. Recently, devices based on MoSi have also been shown to operate as photon detectors which promising results [3], however detailed investigation is necessary to understand how the film growth conditions affect the eventual detector properties.
This study of the MoSi system [4] finds that the superconducting transition temperature reaches a maximum Tc of 7.5 K at a composition of Mo83Si17. The transition temperature and amorphous character can be improved by cooling of the substrate during growth which inhibits formation of a crystalline phase. While X-ray diffraction (XRD) can be used to confirm the absence of significant crystallinity, it is largely insufficient to quantify the extent to which the film is amorphous. We present TEM results which better describe the extent of crystallinity in the films. We observe that for a range of 6 common substrates there is no variation in superconducting transition temperature making MoSi an excellent material for SNSPDs.
The thermal stability of the amorphous films is explored through controlled annealing of the samples. X-Ray Photoelectron Spectroscopy, XRD and transport measurements are used to quantify the extent of crystallisation. In the limits of heating likely to occur during device fabrication, there is very little observed change to the film structure or superconducting properties.
[1] B Baek et al Superconducting a-WxSi1−x nanowire single-photon detector with saturated internal quantum efficiency from visible to 1850 nm APL, 98 25, 251105, 2011
[2] F Marsili et al Detecting single infrared photons with 93% system efficiency Nat. Photonics, 7, 210, 2013
[3] Y Korneeva et al Superconducting single-photon detector made of MoSi film SUST, 27 9, 095012 2014
[4] D. Bosworth et al, Amorphous molybdenum silicon superconducting thin films, AIP Adv. 5, 087106, 2015
10:45 AM - EP7.3.06
Mapping Strain/Pressure with ZnO Nanowire Arrays by Piezotronic and Piezo-Phototronic Effect
Caofeng Pan 1,Zhong Lin Wang 2,Xun Han 1,Rongrong Bao 1
1 Chinese Academy of Sciences Beijing China,2 Georgia Institute of Technology Atlanta United States
Show AbstractEmulation of human senses via electronic means has long been a grand challenge in research of artificial intelligence as well as prosthetics, and is of pivotal importance for developing intelligently accessible and natural interfaces between human/environment and machine. Unlike other senses (seeing, hearing, smelling and tasting), capability of skin for touch sensing remains stubbornly difficult to be mimicked, which necessitates the development of large-scale pressure sensor arrays with high spatial-resolution, high-sensitivity and fast response. In this talk, we present a novel design of nanowire LED arrays, which can be used to directly record the strain distribution by piezo-phototronic effect. This work is published on
Nature Photonics.
1In our previous work, we have demonstrated how the piezo-phototronic effect can be effectively utilized to enhance the emission intensity of an n-ZnO/p-GaN NW LED.
2, 3 The emission light intensity and injection current at a fixed applied voltage has been enhanced by a factor of 17 and 4 after applying a 0.093% compressive strain, respectively. Here, we extend the single NW device to NW LEDs array, for pressure/force sensor arrays for mapping strain with a resolution as high as 2.7 μm. Such sensors are capable of recording spatial profiles of pressure distribution, and the tactile pixel area density of our device array is 6250000/cm
2, which is much higher than the number of tactile sensors in recent reports (~ 6-27/cm
2) and mechanoreceptors embedded in the human fingertip skins (~ 240/cm
2).
When the device is under pressure, the images unambiguously show that the change in LED intensity occurred apparently at the pixels that were being compressed by the molded pattern, while those were off the molded characters showed almost no change in LED intensity. Instead of using the cross-bar electrodes for sequential data output, the pressure image is read out in parallel for all of the pixels at a response and recovery time-resolution of 90 ms. Furthermore, our recent studies achieve such piezo-phototronic effect induced strain mapping in a flexible n-ZnO NWs/p-polymer LEDs array system. This may be a major step toward digital imaging of mechanical signals by optical means, with potential applications in touch pad technology, personalized signatures, bio-imaging and optical MEMS.
1. Pan, C.F.
et al. High-resolution electroluminescent imaging of pressure distribution using a piezoelectric nanowire led array.
Nature Photonics 7, 752-758 (2013).
2. Yang, Q.
et al. Largely enhanced efficiency in zno nanowire/p-polymer hybridized inorganic/organic ultraviolet light-emitting diode by piezo-phototronic effect.
Nano Lett. 13, 607-613 (2013).
3. Yang, Q., Wang, W.H., Xu, S. & Wang, Z.L. Enhancing light emission of zno microwire-based diodes by piezo-phototronic effect.
Nano Lett. 11, 4012-4017 (2011).
EP7.4: III-V Photonics—Lasers on Si
Session Chairs
Wednesday PM, March 30, 2016
PCC North, 200 Level, Room 222 A
11:30 AM - *EP7.4.01
Progress of III-V Membrane Photonic Devices on Si toward On-Chip Interconnection
Nobuhiko Nishiyama 1,Shigehisa Arai 1
1 Tokyo Institute of Technology Tokyo Japan,
Show AbstractProgress of information and communication technology (ICT) is being supported by improvements of LSI performances thanks to miniaturization of CMOS transistors based on the scaling law. However, it is known that LSI performance will soon face their limitation caused by electrical global wires due to transmission loss and cross-talk. To overcome this limitation, an introduction of optical interconnection is a promising solution. In this case, back-end process compatibility should be guaranteed to avoid a conflict with Si CMOS process and design. In addition to this, each photonic component must have ultra-low power consumption of less than 100fJ/bit. As an answer of these requirements, we proposed membrane photonic-integrated-circuits (PICs), which have very thin (200-300 nm) semiconductor layers sandwiched by low index materials such as SiO2 and air. Due to the large index difference, this membrane structure gives strong optical confinement to realize compact devices and low threshold current semiconductor lasers. Up to now, membrane distributed feedback (DFB) lasers, membrane photodiodes (PDs) and wire waveguides as well as simple integrated circuits were demonstrated. In this presentation, device characteristics and their future prospects will be discussed.
A membrane 1.5-µm DFB laser consisting of about 270-nm-thick GaInAsP/InP layers, which included GaInAsP quantum wells and laterally regrown p- and n-type InP cladding layers were bonded on a Si substrate using benzocycrobtene (BCB) bonding. The process temperature was less than 300°C, which is compatible for the back-end process temperature of CMOS circuits. A surface grating gave high index-coupling coefficient of 1500 cm-1 to realize low threshold current operation. By butt-jointed built-in (BJB) regrowth technique, GaInAsP passive waveguide section was integrated and this can reduce the DFB cavity length to several tens microns. Fabricated devices showed typical threshold currents of 1/2 was realized. This means 10 Gbit/s operation can be achieved with less than 1 mA operation current.
Using this BJB regrowth technique, we can also integrate PDs through waveguides for signal distribution. Photo-current signals, which correlated to the output power of lasers, were detected by the integrated PD. By using a discrete PD, a clear eye opening was observed for 10 Gbit/s signals.
By introducing low index materials at the side of waveguides, a III-V wire waveguide structure can be realized similar to Si photonic wire waveguides. We demonstrated low propagation loss of 4 dB/cm with the waveguide size of 420 nm x 150 nm.
12:00 PM - EP7.4.02
Transfer Printing of Thin-Film Microscale GaAs Lasers on Silicon with a Thermally Conductive Interface
Xing Sheng 1,Cedric Robert 2,Brian Corbett 3,John Rogers 4
1 Tsinghua University Beijing China,2 Institut National des Sciences Appliquées de Toulouse Toulouse France3 Tyndall National Institute Cork Ireland4 University of Illinois at Urbana-Champaign Urbana United States
Show AbstractWe exploit microscale, thin-film GaAs lasers integrated onto silicon (Si) substrates via transfer printing, with a thermally conductive interface material for continuous wave (CW) operation at room temperature. The potential for Si based photonics to improve performance in future integrated circuits has created strong demand for efficient on-chip lasers. Strategies that involve separate growth of III-V materials followed by integration on Si offer significant promise. Approaches based on epitaxial liftoff and transfer printing, in particular, have important proven capabilities in this context, with impressive published examples of both edge and surface emitting lasers formed with thin-film, releasable III-V membranes directly bonded to Si. In these schemes, selective removal by wet etching of an epitaxially grown sacrificial layer releases active material structures from the III-V substrate. Soft elastomer stamps serve as non-destructive tools to retrieve these materials and then to deliver them to Si wafers, in array formats in a single step or in a step and repeat fashion. This type of process offers high-speed operation and excellent overlay registration, enabled by controlled van der Waals bonding to the surface of the stamp. A disadvantage of previously reported work in active photonics is that it requires multi-step processing on the Si to complete the lasers, including definition of the cavity, contact metal deposition, etc. In addition, the bonding between III-V materials and Si is most effective with atomically smooth surfaces, thereby creating high levels of sensitivity to parasitic roughness and defects. Here we presents concepts that bypass these challenges, and demonstrates them in strategies for releasing and transfer printing fully formed, functional thin-film microscale gallium arsenide (GaAs) based lasers onto Si substrates where a metallic thin film serves as an adhesive and a thermally conductive interface. Numerical simulations reveal the key considerations in thermal management, with an emphasis on the role of this interface layer. Electrically pumped devices printed on Si exhibit continuous-wave (CW) lasing in the near-infrared range (around 820 nm) at room temperature, with performance comparable to unreleased devices on their native substrates. The results presented here have promise as generalized routes for advanced heterogeneous integration in next-generation electronic and photonic circuits.
12:15 PM - EP7.4.03
Technologies for Transfer Printing of InP Based Etched Facet Lasers
Ruggero Loi 1,James O'Callaghan 1,Cedric Robert 1,Alin Fecioru 2,Antonio Trinidade 2,Christopher Bower 2,Brian Corbett 1
1 Tyndall National Institute Cork Ireland,2 X-Celeprint Cork Ireland
Show AbstractThe data flow over the internet has been growing exponentially over the last 15 years [1]. Silicon photonics is considered to be a solution to overcome the electronic bandwidth and density limits and lead to reduced power consumption in data centers. Silicon on insulator waveguides (SOI) allow the integration of photonics circuits with the electronic infrastructure sending data with light inside thin silicon waveguides on a silicon dioxide confinement layer. A problem with silicon photonics is the lack of a light source and thus strategies to integrate an optical amplifier or laser in a heterogeneous manner are necessary. As silicon is transparent for wavelengths greater than 1 micron, a laser emitting in the telecommunication bands around 1300nm or 1550nm should be used. Here, we use III-V active materials based on AlInGaAs for the realization of lasers which are integrated on silicon substrates by transferring devices to the non-native substrate by means of Micro-Transfer-Printing (μTP) [2-3].
The lasers are pre-fabricated on the native substrate containing the waveguide and p- and n- type materials along with a 1mm thick InGaAs release layers. Inductively coupled plasma etching is used to form the mirrors in a Fabry-Perot cavity with typical length of 500mm while a 2.5mm wide ridge waveguide is produced to guide the light. The etched-facet laser process uses 9 lithography levels with three dedicated to the release and transfer. We have investigated the development of a suitable epitaxial release process with different tether shapes and dimensions to keep the devices anchored during the undercut and the following rinse and dry steps. The undercut process has been investigated using citric acid, iron chloride and phosphoric acid solutions for the InGaAs etch where high selectivity to InP is necessary. By using a transparent elastomeric stamp test laser coupons have been transfer printed from the InP wafer onto Si and glass substrates. In particular, the printing on glass allows the inspection of the adhesion quality between the device and the new substrate where the surface flatness is vital to obtain a proper printing. The electro-optical evaluation of the devices on the native substrate shows laser behavior with parameters in line with the laser geometry used.
REFERENCES
[1] http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/VNI_Hyperconnectivity_WP.html
[2] John Justice, Chris Bower, Matthew Meitl, Marcus B. Mooney, Mark A. Gubbins and Brian Corbett, Wafer-scale integration of group III–V lasers on silicon using transfer printing of epitaxial layers, Nature Photonics, vol. 37, pp. 345-353 (2013).
[3] M.A. Meitl, Z.-T. Zhu, V. Kumar, K.J. Lee, X. Feng, Y.Y. Huang, I. Adesida, R.G. Nuzzo and J.A. Rogers, Transfer Printing by Kinetic Control of Adhesion to an Elastomeric Stamp.. 2006, Nature Materials, Vol. 5, pp. 33-38.
EP7.5: Ge for Si Photonics—GeSn Based Light Sources
Session Chairs
Wednesday PM, March 30, 2016
PCC North, 200 Level, Room 222 A
2:30 PM - *EP7.5.01
Group IV Epitaxial Layers on Si for Photonic Devices
Erich Kasper 2
1 PEK Scientific Consulting Pfaffenhofen Germany,2 University of Stuttgart (retired) Stuttgart Germany,
Show AbstractPlanar silicon (Si) waveguides are the backbone of today’s Si on-chip photonics due to their submicron cross sections and their ability for strongly curved paths. The strong light concentration to the waveguide is caused by the unrivaled high refractive index (n) contrast between the semiconductor core (n=3.5) and the insulator cladding (n=1.5). The Si waveguide is transparent in the near (NIR) and mid infrared (MIR) up from a cutoff wavelength of about 1.1 µm. Monolithic integration of active devices (photodetectors, absorption modulators, light emitting diodes (LED) and lasers) requires heterostructures with low bandgap semiconductors in order to operate in the transparency regime of the Si waveguide.
In this talk we concentrate on group IV heterostructures which would be preferred for Si substrates as compatible material partners. Group IV materials with smaller bandgaps as Si (1.12 eV) are the elements germanium (Ge) and tin (Sn) and their alloys SiGe, GeSn, GeSiSn. The respectable challenges for material science, device physics and characterization are discussed in the talk and the status of epitaxial layer growth is presented and the prospects of photonics/electronics merger are sketched.
Ge which covers the NIR up to 1.55 µm is 4.2% larger than Si, the lattice mismatch between Ge and Sn amounts to a much higher value of 14.7%. Ge rich (about 90% Ge ) GeSn alloys which mark the crossover between the usual indirect group IV semiconductors and the direct semiconductor Sn (Sn itself is metallic at room temperature but has a diamond lattice form at lower temperatures) exhibit a moderate lattice mismatch of 1.5% on Ge but a high lattice mismatch of nearly 6% toward Si substrates. That’s the reason why most device structures are grown as Ge(Si)Sn on a so called virtual Ge substrate (Ge VS) which contains a thin strain relaxed Ge buffer on the Si substrate.
Basic photodetector and LED devices were realized with Ge and GeSn as active materials where even low Sn contents below the indirect/direct crossover extend the infrared range beyond 2µm with prospects for MID up to 5 µm. Electrically stimulated lasers were realized with highly n-doped Ge but needed high injection currents. Promising are efforts to push the active material Ge beyond the indirect/direct crossover by tensile strain or Sn mixing, or better by combining both effects.
Low threshold lasers will additionally need substantial improvements of the electronic material quality with high indirect recombination lifetimes which are now compromised by threading dislocations and additionally in GeSn by high vacancy concentrations.
3:00 PM - EP7.5.02
Group-IV Infrared Light Emitting Diodes on Si
James Gallagher 1,Charutha Senaratne 2,Chi Xu 1,John Kouvetakis 2,Jose Menendez 1
1 Physics Arizona State University Tempe United States,2 Chemistry and Biochemistry Arizona State University Tempe United States
Show AbstractThe pursuit of monolithically integrated light sources on Si has matured significantly in recent years. Here we discuss advances in the development of next-generation group-IV light emitting diodes (LEDs) based on Ge1-ySny (y = 0-0.15) and Ge1-x-ySixSny (x = 0.02-0.10, y = 0.03-0.12) alloys with enhanced performance designs. High-quality crystalline films of these materials are incorporated onto Si platforms via the use of Ge-buffer technologies. In the case of heterojunction pin diodes based on Ge1-ySny and Ge1-x-ySixSny materials, the buffer layers are doped n-type at a level of 2.0×1019 cm-3.
The basic heterojunction pin device architecture is obtained by growing stacks of thick i-layer alloys, 400-700 nm, on the n-Ge/Si(100) substrates followed by the deposition of a highly doped p-type alloy film with p = 1.0-10×1019 cm-3 and thickness 100-200 nm. The growth conditions for the p-type layer are adjusted appropriately to produce films with compositions that ensure pseudomorphic growth of the p-layer on the i-layer while simultaneously providing electrical confinement of the carriers to the i-layer. The fully coherent interface between the i/p films is defect-free, representing a significant improvement over previous state-of-the-art devices that use doped Ge films for contact layers and contain defects at both n/i and i/p interfaces. Electroluminescence (EL) experiments show that as the Sn content of the i-layer is increased, the optical performance improves and both material systems show evidence for a transition direct gap semiconductors. These experiments also indicate that the non-radiative carrier recombination lifetimes are limited by the presence of defects at the bottom interface, hindering optimal device performance.
The carrier recombination lifetimes can be lengthened by removing the mismatch defects at the n/i interface. To accomplish this, the initial design is modified by replacing the n-Ge bottom contact with a highly doped, n = 0.5-2.0×1019 cm-3, n-alloy layer that is grown on an i-Ge/Si(100) substrate. This ensures pseudomorphic growth of the subsequent alloy i- and p-layers. These homostructure pin devices contain no defective interfaces and show remarkably longer carrier recombination lifetimes, as determined by the sharp enhancement of the EL spectra intensity. The work is extended further by fabricating degenerate pn junction LEDs between Ge1-ySny and Ge1-x-ySixSny materials by removing the i-layer from the sample production process. These devices, which represent the basic stack architecture required for an electrically injected laser, show strong EL features corresponding to direct gap emission from the n-layer material that contains an elevated Fermi level due to high doping.
3:15 PM - EP7.5.03
Tuned Ge1-ySny Diode Designs for Investigating the Effect of Strain Relaxation on Electroluminescence
Charutha Senaratne 1,James Gallagher 2,Patrick Wallace 1,John Kouvetakis 1,Jose Menendez 2
1 School of Molecular Sciences Arizona State University Tempe United States,2 Department of Physics Arizona State University Tempe United States
Show AbstractDirect gap Ge1-ySny alloys can be used to fabricate photonic devices that can be integrated onto Si platforms ubiquitous in microelectronics. Significant progress has been made in developing growth techniques for this class of materials, culminating in several reports of device-quality alloys near the indirect-direct crossover composition regime. The devices which make use of these materials have been shown to exhibit electroluminescence (EL) in the mid-infrared. The predominant strategy used to integrate these types of devices onto silicon is through the use of virtual Ge substrates. These consist of thick Ge layers deposited on Si, which serve to minimize the lattice mismatch between the alloy and the substrate, facilitating growth.
For the abovementioned light emitting diodes, the virtual Ge substrate performs an additional function of forming a p- or n-type doped bottom contact in the diode structure. The typical design found for the diodes reported in literature is of an intrinsic Ge1-ySny active layer cladded by two doped Ge layers to make a pin structure. However, this design is disadvantageous for growth of direct gap Ge1-ySny alloys. If the alloy is pseudomorphic to the substrate, the large compressive strain hinders the achievement of direct gap behavior. Conversely, if strain relaxation occurs relative to the substrate, the unavoidable formation of interface defects results in increased non-radiative recombination of the carriers, leading to a degradation of the luminescence intensity.
It has been shown that using Ge1-ySny alloys as the p- and n-type doped contact layers to mitigate interfacial defect formation leads to significant enhancement in EL intensity, even in indirect gap alloys. In this study, this idea was extended to direct gap alloys. The devices produced have a Si/i-Ge/n-Ge0.93Sn0.07/i-Ge0.88Sn0.12/p-Ge0.92Sn0.08 layer structure. In the first stage, a virtual Ge substrate was formed by the deposition Ge on a Si(100) wafer. The Ge1-ySny alloys of the subsequent layers were deposited in a hot wall UHV-CVD reactor using Ge3H8 and SnD4 precursors. The top and bottom layers were doped p- and n-type by use of B2H6 and P(SiH3)3 precursors, respectively. In this design, the top contact layer grows pseudomorphic to the active layer, leading to a defect free i-p interface. There is partial strain relaxation of the intrinsic layer relative to the n-type bottom contact. However, since the lattice mismatch is decreased, the degree of strain relaxation relative to the bottom contact layer is less when compared with devices in which Ge performs this function. By judiciously transitioning the composition of the bottom contact from pure Ge to a Ge1-ySny alloys pseudomorphically matched with the intrinsic layer, we obtain a model system in which to study the effect of defect formation on EL intensity from direct gap Ge1-ySny alloys.
3:30 PM - EP7.5.04
Toward a Direct Bandgap Ge1-xSnx Alloy by Ion Implantation and Pulsed Laser Melting
Tuan Tran 1,Hemi Gandhi 2,David Pastor 2,Lachlan Smillie 1,Michael Aziz 2,Jim Williams 1
1 Department of Electronic Materials Engineering Australian National University Canberra Australia,2 John A. Paulson School of Engineering and Applied Sciences Harvard University Cambridge United States
Show AbstractGe1-xSnx alloys have recently gained widespread attention as being the only group IV semiconductor material that has a truly direct band gap [1]. Attempts at synthesizing the direct bandgap alloy require techniques that provide conditions far from thermodynamic equilibrium, because the required Sn concentration is at least 12 times higher than the equilibrium solubility limit [2]. Most reports on the topic use non-equilibrium growth techniques such as molecular beam epitaxy [3], sputter deposition [4] and chemical vapour deposition[1].
In this contribution, it is shown that ion implantation and pulsed laser melting is another effective method for producing high Sn content Ge1-xSnx material with high crystal quality. To date, up to 6.2% substitutional Sn has been incorporated into Ge by this method. This value is 4 times as high as previous results using a similar concept and close to that required for the direct gap transition [5]. RBS, XRD, Raman spectroscopy and TEM are used to characterise the Sn concentration, crystallinity and defects, strain and other microscopic properties of the material. TEM has shown that the development of a porous structure currently limits an even higher Sn concentration in the lattice but we will report a method to overcome this limitation.
To investigate the effect of soluble Sn content on the alloy band structure, spectroscopic ellipsometry has been used to characterise the optical transitions in the material. Such data show that there is a reduction in the excitonic effect, represented by a lower transition activity at the critical band structure points E1 and E1+Δ1, as well as the shrinkage of the band gap [6]. For the transition (E0) at the direct band valley (Γ), light and heavy hole splitting is observed which could greatly improve the mobility of holes [7]. A significant enhancement of the optical transition in the vicinity of the Γ valley is also observed. Collectively, these data indicate that this material has potential for an efficient, Si compatible photodiode operated at the L band (λ = 1.56 – 1.62 μm) of the optical communication network [8].
References
[1] WirthsS et al., Nat Photon 9, 88 (2015).
[2] R. W. Olesinski and G. J. Abbaschian, Bulletin of Alloy Phase Diagrams 5, 265
[3] G. He and H. A. Atwater, Physical Review Letters 79, 1937 (1997).
[4] S. I. Shah, J. E. Greene, L. L. Abels, Q. Yao, and P. M. Raccah, Journal of Crystal Growth 83, 3 (1987).
[5] K. Gao, S. Prucnal, R. Huebner, C. Baehtz, I. Skorupa, Y. Wang, W. Skorupa, M. Helm, and S. Zhou, Applied Physics Letters 105 (2014).
[6] M. del CastilloMussot and L. J. Sham, Physical Review B 31, 2092 (1985).
[7] J. D. Sau and M. L. Cohen, Physical Review B 75, 045208 (2007).
[8] Y. Ishikawa, K. Wada, J. Liu, D. D. Cannon, H.-C. Luan, J. Michel, and L. C. Kimerling, Journal of Applied Physics 98, 013501 (2005).
EP7.6: Ge for Si Photonics—Novel Devices and Processes
Session Chairs
Wednesday PM, March 30, 2016
PCC North, 200 Level, Room 222 A
4:15 PM - *EP7.6.01
Ge on Si Photonics for Mid-Infrared Sensing Applications
Douglas Paul 1
1 University of Glasgow Glasgow United Kingdom,
Show AbstractMolecular absorption lines in the 3 to 5 µm and 8 to 13 µm mid-infrared (IR) wavelength regions are frequently used to identify chemical and biological gases and liquids. The market presently dictates portable, cheap, single molecular line detectors with low sensitivity for public used or high cost, high sensitivity laboratory based systems that require scientists to operate the tools. Here I will present research aiming to use Ge on Si to develop chemical and biological detectors in the 8 to 13 µm wavelength region as practical systems that could be used for personal healthcare, environmental monitoring or security monitoring.
The bandgap of Ge prevents interband detection in the mid-IR so either intersubband, plasmon or bolometer based detection must be used. I will present Ge on Si based quantum well intersubband photodetectors (QWIPs) operating in the 8 to 13 µm region which can be integrated to create complete sensor systems. The performance of QWIPs improves as the temperature is reduced and so full temperature dependent characterisation will be presented along with how the doping affects the performance.
Heavily doped Ge plasmonic antennas will also be demonstrated which enhance the detection of molecular absorption lines by a factor of 100 compared to straight photodetection. These Ge plasmonic detectors are fully compatible with silicon process foundries unlike many of the metal based plasmonic devices which rely on gold or silver metals. A comparison will be made with the metal plasmonic systems especially with regard to sensitivity and loss. As an example, the Ge plasmonic antennas have been used to detection an explosives simulant and further examples of their use for specific molecular absorption lines will be presented. Finally examples of integrated Ge plasmonic antennas with bolometric detection will be demonstrated as an integrated mid-IR sensor which only requires an external source for a complete system. The performance of the bolometer sensors will be reported along with the measure and potential sensitivity and responsivity.
4:45 PM - EP7.6.02
NIR Laser Annealing Process for Dark Current Suppression in Selectively-Grown Ge Photodiodes on Si
Sho Nagatomo 1,Shinya Kikuta 2,Satohiko Hoshino 2,Yasuhiko Ishikawa 1
1 Department of Materials Engineering Univ of Tokyo Tokyo Japan,2 Technology Development Center Tokyo Electron Limited Nirasaki Japan
Show AbstractGe epitaxial layers on Si are effective for near-infrared (NIR) photodiodes (PDs) in Si photonics. Due to the 4% lattice mismatch between Ge and Si, a high density of threading dislocations (~109 cm-2) is generated in Ge. Post-growth annealing (> ~800°C) is effective for the reduction of dislocation density [1], although such a high-temperature process often causes an intermixing of Ge and Si [2]. SiGe alloys significantly reduce the responsivity in the C/L band (1.53 – 1.62 µm) due to the blue shift in the optical absorption edge. The doping profile in pre-processed Ge/Si devices would be also altered, degrading the device performances. The authors recently proposed an NIR laser annealing as a post-growth annealing of Ge layers [3]. There are mainly two advantages; one is the selective heating of Ge under the NIR irradiation (1.0 - 1.6 µm) due to the optical absorption limited in Ge, while the other one is the short annealing time (< 1 s), minimizing the thermal diffusion of atoms. In this paper, a significant reduction of dark current, corresponding to the reduction of threading dislocation density, is presented for Ge pin PDs on Si using post-growth NIR laser annealing. The laser annealing is also effective even after the formation of pin junction in Ge, reflecting the short annealing time to keep the doping profile.
In experiments, a Ge (600 nm) layer was grown on p+-Si(001) by UHV-CVD, followed by the growth of thin Si cap layer (50 nm). The samples were irradiated with a scanned CW laser light (λ = 1.07 µm, 1 mm in diameter) in air. The laser power and scanning speed were typically 60 W and 10 ms/mm, respectively. The scanning was repeated up to 10 cycles. Phosphorous ions were implanted as n-type dopants in the Si cap and the top region of Ge to form pin structures.
The Ge pin structures showed good rectifying diode characteristics. The dark leakage current was reduced by the laser irradiation/annealing: the dark current density was as large as 70 mA/cm2 at 1 V for the unannealed Ge grown at 600°C, while those for the laser-annealed Ge were reduced to be 20 – 30 mA/cm2, comparable to that for Ge annealed at 800°C in a furnace. The reduced dark current is ascribed to the reduction of threading dislocation density; 5x108 cm-2 for the unannealed one and 2x108 cm-2 for the laser-annealed one, according to the etch-pit counting [1]. It is noted that similar reduction of dark leakage current was obtained for the laser annealing performed after the formation of pin junction in Ge. It is also important that the responsivity was slightly enhanced by the laser annealing particularly for > 1.55 µm, resulting from the increase of tensile strain in Ge. These results suggest that the NIR laser annealing is a promising method as a post-growth annealing for Ge devices on Si.
[1] Luan et al., Appl. Phys. Lett. 75, 2909 (1999). [2] Koester et al., IEEE J. Sel. Top. Quantum Electron. 12, 1489 (2006). [3] Kawamata et al., ECS Trans. 64, 775 (2014).
5:00 PM - EP7.6.03
Silicon-Germanium Engineering and Integration for Photonic and Electronic Applications
Frederic Gardes 1,Callum Littlejohns 2,Milos Nedeljkovic 1,Thalia Dominguez Bucio 1,Nannicha Hattasan 1,Mehdi Banakar 1,Lorenzo Mastronardi 1,Graham Reed 1,Goran Mashanovich 1
1 University of Southampton Southampton United Kingdom,2 NTU Singapore Singapore
Show AbstractFabrication cost, number of channel scalability, low switching energy, and device size are the major prerequisites for future on-chip WDM systems. A number of device geometries have already been demonstrated that fulfil these criteria, at least in part, but combining all of the requirements is still a difficult challenge.
Germanium and silicon-germanium compounds have been demonstrated to be excellent materials for the fabrication of integrated microsystems in CMOS and MEMS. These compounds could fulfil the future requirements for light emission, modulation with devices such as quantum confined Stark effect and Franz Keldysh modulators, and also detectors in photonics integrated devices.
Multiple deposition processes are available for the integration of germanium or silicon-germanium with the silicon-on-insulator platform. The rapid melt growth technique is a viable option for the fabrication of defect free SiGeOI or GeOI, but to date has suffered with the difficulty to obtain a uniform composition of crystalline compound material over a large area. Here we demonstrate the mechanisms involved in obtaining a controlled and uniform SiGe composition of defect free crystalline material, on insulator, over a large area. Moreover we demonstrate the capability to tune the silicon content of the SiGe and have different SiGe content on the same chip. This process represents the fundamental building block required for multilayer compound structures and devices using SiGe and Ge for the simultaneous fabrication of modulators and detectors using a single deposition process, therefore saving cost and energy, and enabling low power high density WDM applications.
5:15 PM - EP7.6.04
Edge Coupled SiGe Mach-Zehnder Modulators
Xiaochen Sun 1,Fuxin Li 1,Zhian Shao 1,Wanping Guo 1,Fei Liu 1,Linghui Jia 1,Ningning Feng 1
1 LaXense Inc. Walnut United States,
Show AbstractBy leveraging Si CMOS technology, Si photonics (SiPh) holds the promise to reduce complexity and cost in many fields. After half century's development of CMOS technology, it becomes highly mature yet highly pragmatic and market driven with many generations (i.e. nodes) coexisting for different products and markets with corresponding volumes. Like MEMS, SiPh cannot be made by a CMOS flow without altering its standard processes and is incompatible with IC's MPW runs, therefore its own volume must justify using a certain node to make economic sense. SiPh developers commonly use CMOS nodes from 130nm to 90nm or 65nm although most key elements do not require resolving dimensions The need for advanced nodes in SiPh mainly comes from the sub-micron Si waveguides which enable high speed performance however introducing difficulties on fiber coupling and dimension tolerance both demanding advanced lithography. The presented work shows using SiGe on SOI enables a dual-waveguide system which can be designed to be either small mode for high efficiency modulation or large mode for easy fiber coupling. And it does not require advanced lithography to process either waveguides.
SiGe, grown SOI, is used as waveguide core cladded by Si and SiO2 at modulator phaser region. By carefully designing SiGe composition and waveguide dimensions, the mode can be sufficiently small and well confined in the SiGe core for high modulation efficiency. The SiGe layer is completely removed out of the phaser region such that the underlying thick Si layer forms a large mode ridge waveguide. The low loss transition from the SiGe to the Si waveguide is achieved by a two-stage tapering mode transformer. Modulators with both lateral and vertical PN junction designs were fabricated by standard Si processes with 0.35µm i-line lithography. A capacitive-loading CPS travelling wave electrode is designed to drive both Mach-Zendner arms in a push-pull configuration to reduce capacitance and driving voltage. The optical and DC measurements have been performed on the fabricated SiGe modulators and show good diode IV behaviors, 5~6dB on-chip optical loss and <1dB waveguide-to-fiber coupling loss thanks to this unique design. The high speed performance is being tested at the time of writing and will be reported in the conference.
5:30 PM - EP7.6.05
Novel Electro-Absorption Modulator with Germanium Fins Evanescently Coupled to Silicon Waveguide
Zhiyuan Jiang 1,Kapil Debnath 1,Graham Reed 2,Shinichi Saito 1
1 Nanoelectronics amp; Nanotechnology Research Group, Faculty of Physical Sciences and Engineering University of Southampton Southampton United Kingdom,2 Optoelectronics Research Centre, Faculty of Physical Sciences and Engineering University of Southampton Southampton United Kingdom
Show AbstractElectro-Absorption (EA) modulator is a promising candidate for the next generation devices in Si photonics with even lower power consumptions for short-reach optical interconnections. However, its relatively higher insertion loss must be addressed, and the efficient coupling from a Si waveguide to EA modulator is a challenging topic. Here, we propose, for the first time, to use novel Ge fin structures for EA modulators. An array of Ge fins evanescently coupled to a Si waveguide is used for the modulation by the Franz-Keldysh effect (FKE). We have designed this new EA modulator using Ge fins based on simulations.
We consider a structure, where Ge fins are connected to heavily doped Si electrodes forming a lateral p-i-n diode. We will apply reverse bias to Ge fins to change the absorption coefficient by FKE. The width and pitch of fins were assumed to be 10-nm and 100-nm, respectively. The single mode Si waveguide with the width of 500-nm and the thickness of 200-nm was located on top of the Ge fins with the oxide of 10-nm in-between.
We assumed that the thickness of fin is to be 70-nm, which guarantees moderate mode overlaps from waveguide to the fin region. We have simulated the modulator operated on wavelength of 1550-nm under two application voltages of 3.3-V and 10-V. We simulated the modulators with different lengths up to 30-μm, and found that both of the insertion loss and the extinction ratio have a perfect exponential dependence on the length of the modulator. Based on this confirmation, we could safely extrapolate to identify the performance of the modulator for the longer length.
Assuming the length of 90-μm, under the drive voltage of 10-V, the proposed device has a remarkably low insertion loss of 0.43-dB and an extinction ratio of 3.31-dB. If we consider the lower drive voltage of 3.3-V, it has the insertion loss of 2.77-dB and extinction ratio of 3.58-dB at the modulator length of 600-μm. We think that the proposed EA modulator with Ge fins will be a promising candidate for compact optical interconnections by Si photonics.
EP7.7: Poster Session: Novel Materials for Integrated Photonics
Session Chairs
Thursday AM, March 31, 2016
Sheraton, Third Level, Phoenix Ballroom
9:00 PM - EP7.7.01
High Speed Silicon Photodetectors with Multiple p-n Junctions in CMOS Technology at 650- and 850-nm Wavelengths
Yue-ming Hsin 2,Yin-Chia Tsai 1,Fang-Ping Chou 1,Chih-Ai Huang 1
1 National Central Univ Jhongli Taiwan,2 University of California, Los Angels Los Angeles United States,1 National Central Univ Jhongli Taiwan
Show AbstractSi photodiodes (PDs) fabricated by standard CMOS technology have been widely examined in optical short-distance communication at 850-nm wavelength because of their low cost and ability to integrate with receivers. As LEDs have become a mature technology, it is promising to use Si PDs in visible light communication (375-780 nm) due to the same vantages of Si. This study investigates a multiple p-n junction Si photodetector fabricated using standard CMOS technology operated at 650- and 850-nm wavelength. Multiple p-n junction photodetectors are formed by stacking several p-n junctions vertically. Due to the wavelength-dependent penetration depth of light in semiconductor, different wavelength characteristics of Si photodetector were evaluated.
In this work a silicon photodetector with triple p-n junctions fabricated in 0.25-µm standard CMOS high voltage mixed signal general purpose process from Taiwan Semiconductor Manufacturing Company (TSMC) was implemented and characterized. Different depths of p-n junction from top to bottom are formed vertically by n-implant / p-well, p-well / n+-buried layer, and n+-buried layer / p-substrate junctions to attain a wavelength-dependent response.
The vertically illuminated measurement involved using a 50-µm multimode fiber with two wavelengths of 650- and 850-nm for responsivity and pulse measurements. The measured photo-currents from deepest p-n junction revealed highest responsivities of 0.21A/W and 0.16 A/W at 850- and 650-nm wavelength, respectively, while biasing at a low value of 1.8 V. It is important to achieve high responsivity in standard bulk CMOS technology at low-bias operation. However, the low responsivity was obtained from shallower p-n junctions due to deep penetration depth of light, but the results from 650-nm wavelength incident light were higher than that from 850-nm wavelength.
The -3 dB bandwidth transferred from pulse measurement showed a bandwidth of 2.4 GHz from deepest p-n junction at 850-nm wavelength and a bandwidth of 1.9 GHz from shallowest p-n junction at 650-nm. The -3 dB bandwidth of 1.9 GHz is among the highest results ever reported in 650-nm wavelength using standard CMOS technology. The proposed multiple p-n junction Si photodetector shows combined excellent performance in 650- and 850-nm wavelength applications.
9:00 PM - EP7.7.02
All Graphene Photodetectors with Ultra-High Sensitivity Assisted by Piezoelectric Substrates
Golam Haider 3,Wei-Heng Shih 4,Yang-Fang Chen 3
1 Engineering and System Sciences National Tsing Hua University Hsinchu Taiwan,2 Institute of Physics Academia Sinica Taipei Taiwan,3 Department of Physics National Taiwan University Taipei Taiwan,4 Department of Materials Science and Engineering Drexel University Philadelphia United States3 Department of Physics National Taiwan University Taipei Taiwan
Show AbstractHybrid quantum dot-graphene photodetectors have recently attracted substantial interest owing to their remarkable performance and low power consumption. However, the performance of the device greatly depends on the interfacial states and photogenerated screening field. As a consequence, the sensitivity is limited and the response time is relatively slow. In order to circumvent these challenges, herein, we have designed a composite graphene and graphene quantum dot (GQD) photodetector on Lead Zirconate Titanate (Pb(Zr0.2Ti0.8)O3) (PZT) substrates to form a ultra-sensitive photodetector over a wide range of illumination power. Under 325 nm UV light illumination, the device shows sensitivity as high as 4.06x109 AW-1, which is 120 times higher than reported sensitivity of same class of devices. Plant derived GQD has broad range of absorptivity and is an excellent candidate for harvesting photons generating electron-hole pairs. Intrinsic electric field from PZT substrate separates photogenerated electron-hole pairs as well as provides the built-in electric field that causes the holes to transfer to the underlying graphene channel. The composite structure of graphene and GQD on PZT substrate produces a simple, stable, and highly sensitive photodetector over a wide range of power with short response time, shows a way to resolve the shortcomings.
9:00 PM - EP7.7.03
Evolution of Structure and Luminescence of InN-on-GaN: From Plains to Pillars
Andrian Kuchuk 1,Morgan Ware 1,Michael Sattler 1,Yuriy Mazur 1,Mourad Benamara 1,Yurii Maidaniuk 1,Petro Lytvyn 2,Gregory Salamo 1
1 Institute for Nanoscience and Engineering, University of Arkansas Fayetteville United States,2 V.Lashkaryov Institute of Semiconductor Physics, National Academy of Sciences of Ukraine Kyiv Ukraine
Show AbstractInN has attracted considerable interest for optoelectronics, high- frequency electronic devices and chemical sensors because of its narrow direct band gap energy, high electron mobility, and high electron saturation velocity. The lack of a lattice-matched substrate and the low dissociation temperature make the growth of high crystalline quality InN films challenging. One-dimensional (1D) structures, such as pillars, however hold the promise of improved crystal quality, as well as the ability to use quantum size effects to control the material properties and create promising novel devices.
In this study, we report on the evolution of structural and optical properties of InN grown on (0001) GaN template by PA-MBE at various temperatures, using a flux ratio of In/N = 1. The change in substrate temperature from 350 to 450oC leads to a transition from 2D planar layer-by-layer growth to 1D pillar formation. X-ray diffraction (XRD) analysis shows that all InN structures are totally relaxed and maintain the same crystallographic orientation as the GaN substrate. The XRD peak widths increase and the intensities decrease as the InN structure changes from 2D to 1D. At the same time, the PL spectra shows the opposite trend, i.e. the PL signal increases as InN structure changes from 2D to 1D. We correlate these observations with changes in the InN structural quality and further investigations at the nano-scale by cross-sectional transmission electron microscopy.
9:00 PM - EP7.7.04
Size and Environment Dependent Structure of ZnS Nanoparticles: A Molecular Dynamics Study
Mohammad Khalkhali 1,Qingxia (Chad) Liu 1,Hongbo Zeng 1,Hao Zhang 1
1 University of Alberta Edmonton Canada,
Show AbstractZnS quantum dots (QDs) have attracted a lot of attention, since they can provide a suitable alternative for cadmium-based QDs, which are known to be highly carcinogenic for living systems. However, the structural stability of noncrystalline ZnS has been shown to be a challenging issue limiting their applicability, since ZnS NPs have the potential to undergo uncontrolled structural changes. Using the molecular dynamics technique, we studied the structural evolution of ZnS nanoparticles with initial zinc-blende and wurtzite crystal structures in bare (vacuum) and hydrated (Water) states.
Simulation results revealed that relaxed configurations of bare ZnS nanoparticles larger than 3 nm consist of three regions: a) crystalline core, b) a distorted network of 4-coordinated atoms environing the crystalline core and c) surface structure entirely made of 3-coordinated atoms. Decreasing the size of the ZnS nanoparticle to 2 nm resulted in disappearance of the crystalline core and further reducing the size caused all atoms to become 3-coordinated. Similarity of surface structures of bare nanoparticles resulted in similar surface properties, which affected their interaction with the surrounding environment. Adsorption of water made crystal structure of nanoparticles more stable. However, surface structure of hydrated nanoparticles showed a great amount of heterogeneity and the aforementioned three-region structure did not form in the hydrated nanoparticles.
Different structural evolution regimes in vacuum and water also affected the dipole moment of nanoparticles, which is the main parameter governing the inter-particle interactions. The distribution of capillary pressure in the bulk of the nanoparticles, as well as the surface stress, being the controlling parameters of vibrational properties of QDs, also evolved differently in the different environments. Results of this study provide a valuable insight into the post-synthesis behavior of ZnS nanoparticles.
9:00 PM - EP7.7.05
Multi-Slit Coupler for Hybrid Silicon Photonics
Ghent Nakamura 1,Hideo Isshiki 1
1 Univ of Electro-Communications Chohu Japan,
Show AbstractIt is common knowledge that silicon is the important material for integrated photonics. However, it cannot be achieved by silicon only and needs hybridization to any other functional materials. Then the deference of reflective index limits design margin of device coupling in the hybrid silicon photonics. In order to overcome this issue, we propose multi-strip coupler (MSC), which can obtain very short coupling length (~10um), low coupling loss (~0.1dB), broadband, and low polarization dependent. Furthermore, MSC is characterized by lower coupling loss with photonic crystal waveguide (PCWG) than that of conventional waveguide. MSC is composed by the periodic narrow strip waveguide with two deferent materials. MSC suppresses the deference of propagation constant owing to the mode conversion, resulting in reduction of the reflection loss. Specifically, multimode established in MSC became the mediator between high order mode in silicon waveguide and low order mode in the coupled waveguide. The main purpose of this research is to analyze this mode matching and find the best structure of MSC. In this research, we use FDTD method for the coupling loss evaluation, and use super cell and equivalent refractive index methods for the propagation constant estimation.
Conventional couplers in integrated photonics usually use a single-mode tapered shape waveguide. The taper couplers are well adopted for the spot size conversion. But they need long conversion length (~300um) because the abrupt changing of propagation constant increases reflection loss. Any optical mode in the input waveguide does not agree with that in the coupled waveguide. On the other hand, MSC can obtain mode matching directly between two different waveguides owing to multimode coupling, resulting in the device length less than 10um.
The input waveguide is 1.6um width and split into the eight narrow strip waveguides at MSC edge. Here the multimode matching is established during the light propagation. MSC is well-match with PCWG compared to conventional waveguides (channel-type, rib-type or strip-type). This is attributed to group velocity anomaly in PCWG. Conventional waveguide cannot have larger propagation constant than that of bulk core material. In contrast, group velocity anomaly generates larger propagation constant in PCWG, and makes propagation smooth. It suggests that the mode matching is dominant in coupling into multimode PCWG.
Meanwhile, it is considered that the multimode in each waveguide and MSC dependents on wavelength and polarization. However, MSC is resistant to the dependence. In FDTD analysis, the coupling loss is settled under 1dB in the wavelength range from 1460um to 1600um independent of TE/TM mode.
In conclusion, MSC is suitable structure for hybrid silicon photonics. Additionally, in hybrid silicon photonics, only PCWG can be fabricated with silicon waveguide at the same time. It enable higher integration and reducing cost of optical integrated circuit.
9:00 PM - EP7.7.06
White Light Emission from Hybrid Structure of Graphene Oxide Quantum Dots/ZnO Nanosheets
Young Jae Park 1,Cuong Tran Viet 2,Kang Bok Ko 1,Chang-Hee Hong 1
1 Chonbuk National Univ Jeonju-si Korea (the Republic of),1 Chonbuk National Univ Jeonju-si Korea (the Republic of),2 Vietnam National University Ho Chi Minh City Viet Nam
Show AbstractWe report herein a new method for hybrid structure of graphene oxide quantum dots(GOQDs)/ZnO nanosheets with strong white luminescence. To fabricate the GOQDs/ZnO nanosheets hybrid structures, GOQDs synthesized by the carbonization of citric acid are simply sprayed onto two-dimensional ZnO nanosheets prepared by hydrothermal growth method on a 5 nm thick Al-coated glass substrate. An increased surface area of ZnO nanosheet makes it possible to capture a large amount of GOQDs during spray-coated process and plays a crucial role in modulating the blue/green emission of GOQDs. Consequently, strong white photoluminescence has been achieved by the integration of the GOQDs and ZnO nanosheets. Our research suggests that white light emission can be generated from this hybrid structure, which can be applied for other promising optoelectronic application.
9:00 PM - EP7.7.07
Compact and Broadband Electro-Optic Modulator in Silicon
Bahawal Haq 1,Mahmoud Rasras 1
1 Masdar Inst of Samp;T Abu Dhabi United Arab Emirates,
Show AbstractDirectional couplers have been used previously for optical switching and modulation in LiNbO3, GaAs, and electro-optic polymers. A thermo-optic switch in SOI platform has also been demonstrated. Directional coupler based modulators can be designed to be compact, efficient and to attain higher speeds.
We propose a compact electro-optic modulator in silicon, based on plasma dispersion effect [1] in a wavelength insensitive directional coupler [2]. The device consists of three directional coupler stages separated by constant passive waveguide phase delay section. The first and last coupling stages are passive while the middle one has an active high-speed p+-i-n+ junction embedded in the coupling region. There is a continuous transfer of optical energy between the waveguides in the coupling region. By injecting the charge through an active p+-i-n+ junction, a large change in the absorption coefficient and the effective index of the waveguides can be produced, resulting in a significant extinction of the optical power in both waveguides. Rib waveguides are 0.5 μm in width and 0.22 μm in height. The slab thickness is 90 nm. The doping concentration of highly doped regions are 2x1020 cm-3 and they are placed 0.4 μm apart from rib waveguides. To increase the modulator efficiency, a small part of rib waveguides (0.1 μm) and adjacent slab regions are p-doped with 9x1017 cm-3 acceptor doping and n-doped with 3x1017 cm -3 donor doping. Using nonlinear optimization technique, the coupling strength of each stage and the fixed phase delays are optimized to produce a broadband spectral response. Furthermore, the doping profile of the n- and p-doped regions are optimized to yield high-speed response while keeping the insertion loss very low (0.1dB). We studied the effect of various physical parameters by investigating the electrical and optical response of the device. We show that by tuning the effective index and the loss coefficient in the active coupling region a compact and broadband modulator with a high extinction ratio can be designed. The modulator can be designed to be less than 100 µm in length and with an extinction ratio of greater than 20 dB. The device has a spectral bandwidth of 107 nm. The modulator has a high modulation efficiency with DC Vπ L of 0.034 V-cm. Using pre-emphasized driving signal, the modulator can support moderate data rates.
1. R. A. Soref and B. R. Bennet, ”Electrooptical Effects in Silicon,” IEEE J. Quantum Electron. 23, 123-129 (1987).
2. K. Jinguji, et al., ”Mach-Zehnder interferometer type optical waveguide coupler with wavelength-flattened coupling ratio,” Electron. Lett., vol. 26, pp. 1326-1327,1990.
9:00 PM - EP7.7.08
Resonant Vibrational Excitation of Ammonia in the Synthesis of Gallium Nitride
Yunshen Zhou 1,Hossein Rabiee Golgir 1,Kamran Keramatnejad 1,Wei Xiong 1,Dawei Li 1,Mengmeng Wang 1,Lijia Jiang 1,Xi Huang 1,Yong-Feng Lu 1
1 Univ of Nebraska-Lincoln Lincoln United States,
Show AbstractGallium nitride (GaN), a direct bandgap III-V semiconductor, is a multifunctional material attractive to a broad range of applications due to its unique material properties. The wide bandgap (3.4 eV) endows its distinctive capabilities in optoelectronics, such as light-emitting diodes and blue / violet laser diodes. Its inertness to ionizing radiation makes it a suitable material for outer-space applications. The high electron mobility and breakdown voltage enable GaN high electron mobility transistors (HEMTs) for power electronics and high-power-high-frequency devices. The emerging applications challenge GaN synthetic methods of high efficiency, reliability and cost effectiveness. Metal organic chemical vapor deposition (MOCVD) is a standard method for growing GaN thin films. However, MOCVD of GaN suffers from relatively low growth rate (around 2 μm/h) and high reaction temperature (around 1000 Celsius), which is required for cracking ammonia (NH3) and overcoming energy barriers. However, it also leads to negative impacts, including high defect concentrations, GaN decomposition, and nitrogen loss. Therefore, alternative approaches are desired to substantially reduce reaction temperature and improve GaN growth rates.
In this study, we report the influence of resonant vibrational excitation of NH3 in growing GaN via a laser-assisted MOCVD (LMOCVD) method. By resonantly exciting the NH-wagging mode (ν2) of NH3 molecules at 9.219 µm, the highest growth rate of 84 µm/h was achieved at 750 Celsius. The lowest growth temperature was realized at 250 Celsius. According to optical emission spectroscopic investigation, resonant excitation of NH3 leads to effective NH3 decomposition and promotes the supply of active nitrogen species in GaN formation.
9:00 PM - EP7.7.09
Advances in Crystalline ZnSe Optical Fibers Deposited via High Pressure Chemical Vapor Deposition
Stephen Aro 1,Michael Coco 1,Angela Leone 1,Justin Sparks 1,Pier Sazio 2,Venkatraman Gopalan 1,John Badding 1
1 Pennsylvania State University State College United States,2 University of Southampton Southampton United Kingdom
Show AbstractChemical vapor deposition (CVD) is an established technique which is well suited to fabricating high quality semiconductor films. However uniform conformal coating of templates becomes increasingly difficult as template dimensions decrease below the mean-free path of the precursor molecules within the reaction chamber. High-pressure chemical vapor deposition (HPCVD) uses carrier gas pressures of hundreds of atmospheres to facilitate effective mass transport of reactive semiconductor precursors into ultra-high aspect ratio templates which are otherwise incompatible with CVD.1 In this way ultra-high-aspect crystalline semiconductor optical fibers can be fabricated. Semiconductor fiber optics are of great interest because they allow for the exploitation of the optoelectronic properties of semiconductor materials within all-fiber devices. In this work an established HPCVD synthesis was used to create ultra-high aspect ratio ZnSe fiber optics with centimeter lengths and axial dimensions of microns.2 Process improvements such as ultra-high vacuum precursor preparation have led to decreased optical losses in these structures as compared to previously reported values.3 Multimode waveguide losses as low as 0.2dB/cm in the near infrared have been demonstrated in 15µm core diameter fibers. While the as-deposited structures are functional as waveguides, refinement of the structure should lead to improvements in the optical damage threshold and light transmission. Constraints on the precursors which may be used in the HPCVD process result in a nanoscale central void remaining in the ZnSe waveguides. Although small this void acts as a defect site decreasing the optical damage threshold which is still an impressive 5MW/cm2. This threshold should increase if the central void can be removed. Additionally it has been demonstrated that the dominant optical loss mechanism in the as-deposited structures is scattering loss from grain boundaries within the material.3 Increasing the grain size via post-processing is therefore of great interest. However simple thermal annealing leads to large strains and subsequent cracking within the ZnSe core due to the large disparity in thermal properties between the ZnSe core and the silica glass cladding. The need for a post processing method utilizing decreased temperatures has led to the development of microscale chemical vapor transport (CVT) techniques. CVT has been demonstrated to improve both material crystallinity and optimize morphology at significantly reduced temperatures (650oC) as compared to simple thermal annealing (1000oC). Furthermore a more thermally compatible cladding glass has been developed to aid in the CVT processing and further minimize the thermally induced strain within the ZnSe fiber cores. As CVT techniques are further optimized high power solid core fiber optic waveguides with transparency across the infrared could be realized.
1.Science 2006 311 1583
2.Adv Funct Mater 2013 23 1647
3.Adv Mater 2011 23 1647
9:00 PM - EP7.7.10
Study of InAs QD on Pyramidal GaAs Structure by Confocal Raman Spectroscopy
Leon Hamui 2,Alberto Piedra 1,Yenny Casallas 1,Dagoberto Cardona 1,Salvador Gallardo 1,Maximo Lopez-Lopez 1
1 CINVESTAV Mexico City Mexico,2 UNAM Mexico Mexico,1 CINVESTAV Mexico City Mexico
Show AbstractWe studied a InAs/GaAs:Mg/GaAs quantum dot nanostructures grown by Molecular Beam Epitaxy on GaAs (001) oriented substrate. A pyramidal structure has been obtained for several deposition conditions and varying the temperature it has been observed a change on the density of these pyramids. By SIMS it has been observed that the InAs quantum dots are preferentially deposited on the pyramidal structure. Complementary to direct optical imaging by microscope, Raman Spectroscopy which is widely applied to study nanostructures has been used to determine the different phonon modes presented. In order to study the pyramidal effect on the QD deposition, Confocal Raman Spectroscopy analysis of the surface allows profiling the Raman intensity of the different phonon modes over a scan area. Raman spectroscopy allows not only structural parameters such as superlattice periodicity and layer composition to be determined, but also provides information of the QD size, shape and strain relaxation. Moreover, we have been able to scan an area with depth profiles giving as a result a stack used to obtain 3D (tomographic) spatial information of the sample. Therefore we were able to determine the composition and strain of the pyramids via 3D imaging compared to the flat film area and to understand the favorable localized deposition of the InAs QD on the pyramids. Although, it has been observed by Raman Spectroscopy that there is an increase of the InAs QD signal at the top of the pyramid and the GaAs TO signal is downshifted due to the Mg doping which is more intense within the pyramid area. The confocal area image was colored for the different phonon modes to be able to study the growth of the pyramidal structure and relate it to the deposition temperature.
9:00 PM - EP7.7.11
Composition Graded InGaAsP Alloys Nanowires Grown by Dual Gradient Chemical Vapor Deposition
Seyed Ebrahim Hashemi Amiri 2,Sunay Turkdogan 3,Fan Fan 1,Yueyang Yu 1,Zhicheng Liu 1,Praneeth Ranga 1,Cun-Zheng Ning 1
1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States,2 Department of Chemistry Arizona State University Tempe United States,1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States,3 Electronic and Communication Engineering University of Yalova Yalova Turkey1 School of Electrical, Computer and Energy Engineering Arizona State University Tempe United States
Show Abstract
Growth of composition graded semiconductor alloy nanowires using the dual gradient method [1] has produced an interesting material platform for developing monolithic solar cells for dispersive concentration photovoltaic systems [2]. At the same time, such growth also provides a unique means to study semiconductor alloys in high quality single crystal form and in a wider composition range that is otherwise impossible to grow. We have in the past studied such growth and the resulting materials using ZnCdSeS in its full composition range [1,3]. For many optoelectronic device applications, III-V material systems such as InGaAsP provide more advantages in terms of material properties and existing rich knowledge in device fabrication based on these materials. InGaAsP system is able to cover a large range of bandgaps of (0.35, 2.25) eV, appealing for many applications. Thus similar study of III-V materials is of great interest.
In this paper, we will show how dual gradient method provides a combinatorial chemistry type of platform for growing the composition graded ternary and quaternary nanowire alloys of (In,Ga)(As,P) material system by tuning the supersaturation rate in an Au-catalyzed Vapor-Liquid-Solid (VLS) growth process on a single substrate. More specifically, we have studied the chemical vapor deposition (CVD) growth of InGaAsP with emphasis on the effects of vastly different sublimation rates of the associated elements. Due to the higher sublimation rates of As and P than those for In and Ga, the stoichiometry of the resulting alloys is greatly influenced. To address this issue, different growth strategies are experimented to compensate the deficiency of As and P. The dual gradient approach provides an especially effective tool for rapidly identifying growth conditions in terms of adjusting the temperature and source element gradients. . As a result the bandgap gradually changes on a single substrate for different alloy system. The effect of thickness of Au catalyst, source and substrate temperature and over flux of As and P were studied systematically. The result of growth for GaP rich-sides showed that both the size of Au particle and temperature significantly diminish the formation of stable oxide species such as Ga2O3 and GaPO4, leading to higher quality growth of desired InGaAsP materials. Furthermore, we will present growth of Indium-catalyzed growth of InGaAsP alloys using our growth methodology. We believe that growth of III-V nanowire alloys in a wide range of band gaps within a single substrate can pave the way for fabrication of variety of new multifunctional optoelectronic devices such as multispectral photodetectors, broad band tunable LEDs and laser on a single chip, multispectral photodetectors and high-efficiency solar cells.
References
[1] A.Pan et al., ACS. Nano, 4 (2), pp 671–680, 2010.
[2] Caselli, Derek, et al. Nano letters 14.10 (2014): 5772-5779.
[3] X. Zhaung et al., Advanced Materials 24.1 (2012): 13-33.
9:00 PM - EP7.7.12
Electric Field Induced Structural Colour Tuning of a Silver/Titanium Oxide Nanoparticle One Dimensional Photonic Crystal
Eduardo Aluicio-Sarduy 2,Simone Callegari 2,Diana Figueroa del Valle 1,Ilka Kriegel 1,Francesco Scotognella 2,Guglielmo Lanzani 1
1 Istituto Italiano di Tecnologia (IIT) Milano Italy,2 Politecnico di Milano (PoliMi) Milano Italy,2 Politecnico di Milano (PoliMi) Milano Italy1 Istituto Italiano di Tecnologia (IIT) Milano Italy
Show AbstractPhotonic crystals are promising materials for optics and photonics, owing to the fact that the modulation of the refractive index in these crystals allows a spectral control of the light transmission. A drawback of the photonic crystals is that the tuning of the modulation of the refractive index, due to the alternation of different materials, needs the use of materials that can swell and shrink upon a chemical action or the use of non-linear materials.
Recently, the possibility to tune the localized surface plasmon resonance of gold nanoparticles with the electric field has been demonstrated.
In our work, we have demonstrated that a change of the plasmon resonance in a silver nanoparticle layer, induced by an external electric field, gives rise to a change in the dielectric function, resulting in a shift of the photonic band gap of a silver nanoparticle/titanium dioxide nanoparticle photonic crystal. We fabricated the photonic crystal between two indium tin oxide electrodes and we clearly observed a blue shift of the photonic band gap, upon electric field, together with the blue shift of the silver plasmon resonance. These results indicate that the Silver/TiO2 nanoparticle based photonic crystals could be used as electro-optic switches.